HAL Id: hal-02912034
https://hal.inrae.fr/hal-02912034
Submitted on 5 Aug 2020
HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.
quatre’ scenario
Jérôme Artus, Isabelle Hue, Hervé Acloque
To cite this version:
Jérôme Artus, Isabelle Hue, Hervé Acloque. Preimplantation development in ungulates: a ’ménage à quatre’ scenario. Reproduction, BioScientifica, 2020, �10.1530/REP-19-0348�. �hal-02912034�
© 2020 Society for Reproduction and Fertility https://doi.org/10.1530/REP -19-0348
ISSN 1470–1626 (paper) 1741–7899 (online) Online version via https://rep.bioscientifica.com
REPRODUCTION
REVIEWPreimplantation development in ungulates: a ‘ménage à quatre’
scenario
Jérôme Artus
1,2, Isabelle Hue
3and Hervé Acloque
41UMR935, Inserm, Villejuif, France, 2Université Paris Saclay, Université Paris Sud, Faculty of Medicine, Le Kremlin Bicêtre, France, 3UMR BDR, INRA, ENVA, Université Paris Saclay, Jouy-en-Josas, France and 4GABI, INRA, AgroParisTech, Université Paris Saclay, Jouy-en-Josas, France
Correspondence should be addressed to H Acloque; Email: herve.acloque@inra.fr
Abstract
In ungulates, early embryonic development differs dramatically from that of mice and humans and is characterized by an extended period of pre- and peri-implantation development in utero. After hatching from the zona pellucida, the ungulate blastocyst will stay free in the uterus for many days before implanting within the uterine wall. During this protracted peri-implantation period, an intimate dialog between the embryo and the uterus is established through a complex series of paracrine signals. The blastocyst elongates, leading to extreme growth of extra-embryonic tissues, and at the same time, the inner cell mass moves up into the trophoblast and evolves into the embryonic disc, which is directly exposed to molecules present in the uterine fluids. In the peri- implantation period, uterine glands secrete a wide range of molecules, including enzymes, growth factors, adhesion proteins, cytokines, hormones, and nutrients like amino and fatty acids, which are collectively referred to as histotroph. The identification, role, and effects of these secretions on the biology of the conceptus are still being described; however, the studies that have been conducted to date have demonstrated that histotroph is essential for embryonic development and serves a critical function during the pre- and peri implantation periods. Here, we present an overview of current knowledge on the molecular dialogue among embryonic, extraembryonic, and maternal tissues prior to implantation. Taken together, the body of work described here demonstrates the extent to which this dialog enables the coordination of the development of the conceptus with respect to the establishment of embryonic and extra-embryonic tissues as well as in preparation for implantation.
Reproduction (2020) 159 R151–R172
Introduction
In most mammals, the embryo develops in a dedicated environment, the uterus, which serves both nutritive and protective functions. As a consequence, the main activity of the first period of embryonic development is the production of extraembryonic tissues which are essential for placentation and for the survival of the embryo in utero. Early mammalian embryogenesis has been extensively studied in rodents and primates, with the mouse as the main reference species of the last several decades. These studies have identified the key principles that govern early development including morphogenesis, fate commitment, and pluripotency (Artus & Hadjantonakis 2012). From fertilization to the blastocyst stage, a conserved succession of morphogenetic events is observed in all mammalian species and occurs at a similar rate, lasting from 4.5 days to 7–8 days. During this period, highly conserved developmental steps result in the differentiation and segregation of embryonic from extraembryonic tissues (although these tissues sometimes remain intermingled
within the same cell layer, as in marsupials). At the end of this process, the resulting blastocyst is composed of three distinct cell populations: (i) the epiblast, which is pluripotent and forms the embryo proper, and two sets of extraembryonic tissues that are necessary for specialized interactions with the maternal uterus, (ii) the trophoblast or trophectoderm, an extra-embryonic layer that contributes to the fetal placenta, and (iii) the hypoblast or primitive endoderm, an extra-embryonic layer that gives rise to the visceral and parietal hypoblast (Acloque et al.
2012). At this point, different developmental strategies evolved in different mammalian species, leading either to implantation (as in mice or humans) or to an extended period of pre-implantation development frequently associated with extreme growth of extra- embryonic tissues and the initiation of gastrulation and morphogenesis (at the end of this process). This specific developmental window has been described in many ungulate species (Hue et al. 2012). During this period, the conceptus remains in contact with uterine fluids, through the external layers composed of the trophoblast and the epiblast.
Downloaded from Bioscientifica.com at 08/05/2020 08:01:07AM via Inra
To date, little is known of the complex molecular dialog that occurs between the maternal uterus and the first three main cell populations of the conceptus. This dialog and its molecular dynamics are key to understanding the coordination of numerous crucial developmental processes, including the elongation of extra-embryonic tissues, visceral hypoblast formation, epiblast priming, initiation of gastrulation, primordial germ cell formation, left–right symmetry specification, and finally conceptus implantation. The aim of this review is to summarize the current state of knowledge of this ‘ménage à quatre’. We also examine how pluripotency is controlled in some ungulate species and why it remains so complicated to reproduce in vitro this amazing in vivo process.
Preimplantation development in ungulates and early lineage specification
Timing and early morphogenesis of an ungulate conceptus
During its journey from the oviduct to the uterus, the embryo undergoes a series of cleavage and morphogenetic events whose dynamics and timing vary depending on the species in question. The first noticable change in the embryo is a flattening of the blastomeres, giving it the appearance of a mulberry and thus the name morula. This process, referred to as compaction, results from an increase in intercellular adhesion and the acquisition of cell polarity. Compaction is also closely associated with the differentiation of the inner cell mass (ICM) from the trophectoderm (TE) in the mouse. In ungulates, this occurs around day 5, at the 16–32-cell stage in bovines (Van Soom et al. 1997), pigs (Reima et al. 1993), and sheep (Bindon 1971). The second major morphogenetic event is the formation of the blastocyst, characterized by a fluid-filled cavity known as the blastocoel. This process of cavitation occurs around day 6–7 in bovines (Van Soom et al. 1997) and day 7–8 in pigs (Reima et al. 1993) and sheep (Bindon 1971). The timing of these developmental processes is quite similar in all mammalian species – activation of the embryonic genome occurs between the 5-cell and the 16-cell stage – and the main molecular players and signaling pathways described in rodents and primates are also conserved in ungulates (Piliszek & Madeja 2018). The main differences that have been described among species so far mostly relate to variations in functional dynamics, which will be detailed later.
In most ungulates, after the blastocyst hatches from the zona pellucida, it remains free in the uterine tract for several days before implanting within the uterine wall.
This protracted peri-implantation period distinguishes ungulate development from that of rodents and primates and is further characterized by the establishment of an intimate dialog between the embryo and the uterus through a complex series of paracrine signals (detailed
in ‘Maternal influence on conceptus development’
section). During this period, which varies widely among species, the spherical blastocyst elongates into a tubular and filamentous form. In pigs, this process starts on day 10 and transforms a sphere 0.5–1 mm in diameter into a filamentous blastocyst 1000 mm long by day 16. Interestingly, elongation of the blastocyst seems to be initially driven by cellular hypertrophy rather than cellular hyperplasia (Geisert et al. 1982, reviewed in Bazer & Johnson 2014). As Rauber’s layer disappears, the ICM moves up into the trophectoderm and evolves into an embryonic disc which eventually starts to gastrulate (Fléchon et al. 2004, van Leeuwen et al. 2015).
The terminology used to describe the different layers that make up an ungulate blastocyst can sometimes be misleading, mostly because either similar terms are assigned to inequivalent cell layers or different terms are used to describe the same cell layer. As emphasized in Fig. 1 and according to Pfeffer et al. (2017), we propose the following terminology. The early blastocyst is composed of the trophoblast (TB) and the inner cell mass (ICM). After hatching, the three main cell populations of the blastocyst are the trophoblast (TB), the epiblast (EPI), and the hypoblast (HYPO). At the spherical stage, two main territories can be observed: (1) the embryonic disc, which includes the EPI, the visceral hypoblast (VH), and the disappearing Rauber’s layer (the equivalent of the polar trophectoderm in mice), and (2) the extra-embryonic parietal layers which include TB (external layer) and the parietal hypoblast (PH, internal layer). At the ovoid stage, TB can be subdivided into two types: the adjacent mural trophoblast (AMTB), located in close contact with the EPI, and the mural trophoblast (MTB) overlying the PH. As development progresses, the EPI also starts to differentiate through the formation of the posterior marginal zone of the epiblast (PMZE) which delineates the antero-posterior axis. The VH can also be subdivided into two regions: the anterior visceral hypoblast (AVH) overlying the EPI and the VH underlying the PMZE. We will use this terminology for this review.
TB/ICM specification
The first specifications in cell lineage in the mouse integrate various inputs including cell positioning, cell polarity, mechanical tensions, and metabolic constraints (Kim et al. 2018, White et al. 2018). These signals seem to be integrated by the Hippo signaling pathway which in turn regulates the expression of key lineage transcription factors (TFs); these establish a stable gene regulatory network that controls cell lineage specification and subsequent cell fate determination (Nishioka et al. 2009).
Cell polarization at the time of compaction is one of the key determinants and asymmetric cell division generates polar and apolar cells on the basis of inheritance of the apical domain (polar cell) (Johnson & Ziomek 1981). While apolar cells tend to be located inside the
Figure 1 Schematic representation of blastocyst stages in pigs and bovines. The terminology used to identify the different cell layers and tissues is explained for each stage. The timeline depicts, for each species (pig, bovine, mouse, and human), the developmental time needed to reach a given blastocyst stage.
developing embryo, polar cells remain outside, facing the external environment. The apical domain sequesters some components of the Hippo signaling pathway, including AMOT, an activator of LATS kinases, so that the pathway is inactive in polar cells (Hirate et al. 2013, Leung & Zernicka-Goetz 2013). This absence of Hippo signaling activity leads to the nuclear localization of Yes-associated protein (YAP) which, by interacting with the transcription factor TEAD4, drives the expression of TB-specific genes. Conversely, the active Hippo pathway in apolar cells leads to the phosphorylation of YAP and its consequent cytoplasmic localization, which prevents its interaction with TEAD4 and eventually blocks the commitment of these cells toward the TB lineage.
In other mammals, the Hippo signaling pathway is likely to be a similarly major driver of TB/ICM specification. Comparable signaling machinery has been described in pigs (Emura et al. 2016, Liu et al. 2018) and bovines (Home et al. 2012, Ozawa et al. 2012, Sakurai et al. 2016). As in the mouse, ROCK inhibition in pig embryos promotes Hippo signaling activity, suppressing CDX2 (Caudal type Homeobox 2) expression, a key transcription factor for TB specification and maintenance (Kono et al. 2014, Liu et al. 2018). However, the reverse effect was reported in bovines, which suggests a certain degree of variability among species (Negrón- Pérez et al. 2018). Despite this difference, alterations in YAP or AMOT expression affect bovine blastocyst
formation and the number of CDX2-positive cells is negatively correlated with TEAD4 expression level (Sakurai et al. 2016, Negrón-Pérez et al. 2018). While the functional link between cell polarization, activity of the Hippo signaling pathway, and CDX2 expression is conserved between mice and pigs and, by extension, other mammals, the molecular kinetic for TB lineage determination differs among species from that described in the mouse (Liu et al. 2018).
In the mouse, the Hippo pathway regulates key TB-specific genes as Cdx2, Gata2, and Gata3, which encode transcription factors that are involved in setting up the TB genetic program (Ralston et al.
2010). In particular, CDX2 appears to be critical for TB maintenance rather than TB specification (Strumpf et al.
2005). Similarly, in pigs, CDX2 is necessary for TB cell proliferation and the maintenance of cell polarity (Bou et al. 2017), and in bovines, for TB maintenance at later embryonic stages (Berg et al. 2011, Goissis & Cibelli 2014, Sakurai et al. 2016).
GATA2 and GATA3 (GATA-binding proteins 2 and 3) are also involved in mouse TB maintenance (Ray et al. 2009, Ralston et al. 2010, Home et al. 2017).
Interestingly, these GATA-TFs also regulate TB markers in the bovine trophoblast CT-1 cell line, which suggests a conserved role in TB biology. However, they also act on species-specific targets and this role is exemplified by their regulation of Interferon tau expression in bovines,
which is critical to pregnancy recognition in ruminants (Bai et al. 2011). However, to date, no analysis has been reported that assesses the function of GATA2 and GATA3 in the early embryo in ungulates. Additional TFs have also been identified as regulators of TB physiology by Pfeffer (2018) and Piliszek and Madeja (2018); these studies highlighted both common and species-specific TB regulators.
In addition to Hippo signaling, the specification of TB/
ICM cell lineages appears to be regulated by multiple other pathways, including Wnt signaling. Indeed, at least two studies have examined the effect of modulation of Wnt signaling activity on early bovine embryonic development (Denicol et al. 2013, Madeja et al. 2015).
Denicol et al. (2013) reported that activation of Wnt reduced numbers of TB and ICM cells, while treatment with DKK1, a Wnt inhibitor, had no major effect on blastocyst formation. In contrast, Madeja et al. (2015) showed that Wnt activation upregulated pluripotent markers at the expense of TB markers. In addition, Wnt signaling played a role in trophoblast stem cell maintenance in vitro, likely through the regulation of YAP/TAZ activity (Wang et al. 2019). This study suggested the presence of essential cross-talk between Wnt and Hippo signaling pathways. Interestingly, DKK1 was reported to regulate blastocyst elongation during the peri-implantation period (Tribulo et al. 2019). Taken together, these studies highlight the important role(s) of the Wnt pathway in TB cell biology, specifically in promoting self-renewal and cell differentiation.
Another pathway that seems to be involved in early ungulate embryonic development is JAK/STAT signaling.
For example, STAT5 is expressed during early bovine embryogenesis (Flisikowski et al. 2015) and some STAT1 and STAT3 SNPs were found to be associated with improved embryonic survival (Khatib et al. 2009).
Lastly, inhibition of JAK1/2 affects formation of the ICM but not the TB (Meng et al. 2015). Several questions still remain to be answered, including which upstream signal(s) regulate(s) JAK/STAT, which genes are targeted, and whether cross-talk exists with Hippo signaling.
An aspect of ungulate development that appears to be unique relates to the fate of the TB. At the hatched blastocyst stage, TB can be subdivided into two spatially distinct tissues: TB that covers the blastocoel cavity, and the polar trophoblast, also known as Rauber’s layer (RL). In rodents and primates, the polar trophoblast contributes to the post-implantation development of the fetal part of the placenta; in ungulates, instead, this function is fulfilled by the mural trophoblast. Indeed, one characteristic of ungulate early embryonic development is the disappearance of the RL, occurring around days 9–11 in pigs (Sun et al. 2015), days 10–12 in horses (Enders et al. 1988), and day 14 in bovines (Maddox- Hyttel et al. 2003, van Leeuwen et al. 2015). After RL disintegration, the EPI is completely devoid of TB and thus directly exposed to uterine fluids. RL disappearance
is thought to occur via apoptosis rather than as the result of a difference in proliferative rate with the EPI (Enders et al. 1988, Maddox-Hyttel et al. 2003). The role of RL loss during early embryonic development is still under debate, but it has been linked to the formation of the anterior visceral hypoblast and the emergence of the primitive streak (van Leeuwen et al. 2015).
EPI/HYPO specification
The ICM eventually differentiates into two distinct cell lineages, EPI and HYPO. In the mouse, their formation results from a sequence of events that was originally studied through the dynamic expression of EPI- and HYPO-specific TFs (Artus & Chazaud 2014). NANOG (EPI) and GATA6 (HYPO) are initially expressed in all blastomeres at the morula stage (Plusa et al. 2008). Their expression is progressively restricted and eventually becomes mutually exclusive, so that ICM cells express only the EPI or the HYPO genetic program. By the mid- blastocyst stage, the ICM is a mixture of EPI and HYPO progenitors organized in a salt-and-pepper pattern (Chazaud et al. 2006). Once specified, EPI and HYPO cells reorganize so that HYPO cells lie in contact with the blastocoel cavity, encapsulating the EPI. This cell- sorting process is likely to involve cell adhesion, active migration, positioning information, epithelialization, and selective apoptosis of mispositioned cells (Gerbe et al. 2008, Plusa et al. 2008, Meilhac et al. 2009).
Whether this succession of steps is conserved in other mammals remains to be clarified. However, based on expression studies, this pattern seems to be replicated in ungulates, although with different kinetics, as reported from studies of bovines (Khan et al. 2012, Kuijk et al.
2012), pigs (Ramos-Ibeas et al. 2019), and horses (Choi et al. 2015).
In the mouse, the specification of EPI and HYPO cells is accompanied by a progressive loss of plasticity (Chazaud et al. 2006, Grabarek et al. 2012). Interestingly, HYPO formation is associated with the sequential expression of different markers, which is assumed to represent the activation of a lineage-specific genetic program. These markers are, successively, GATA6 (8-cell), PDGFRα (16-cell), SOX17 (32-cell), GATA4 (64-cell), and SOX7 (sorted PrE cells) (Plusa et al. 2008, Artus et al. 2011).
The role of these factors remains to be determined in other mammals. In this situation, cell fate specification of a bi-potential ICM cell can be viewed as a process in which one genetic program is shut down while the other is maintained. This mechanism must be tightly regulated temporally and perhaps spatially and requires the regulation of transcriptional activity as well as of protein stability (Bessonnard et al. 2017).
In mammals, EPI/HYPO formation seems to be strictly regulated by receptor tyrosine kinase (RTK) signaling.
In rodents in particular, EPI/HYPO specification is controlled by fibroblast growth factor (FGF) signaling,
and this process has been largely characterized through gain- and loss-of-function experiments (Chazaud et al.
2006, Nichols et al. 2009, Yamanaka et al. 2010, Kang et al. 2013, 2017, Krawchuk et al. 2013, Molotkov et al. 2017). There is abundant evidence, however, to suggest that the role of FGF signaling in EPI/HYPO cell- fate decisions is not confined to rodents. In bovines, the inhibition of specific FGF signal transducers like the mitogen-activated protein kinase (MEK) and FGF receptor (FGFR) has different outputs. Indeed, inhibition of MEK but not FGFR biases cell-fate decisions toward EPI, while the addition of FGF4 promotes development into HYPO (Kuijk et al. 2012). Similar observations were reported in ovine blastocysts (Moradi et al. 2015).
However, in humans, MEK inhibition does not affect EPI/HYPO specification (Kuijk et al. 2012, Roode et al.
2012). These observations suggest that (i) other input signals regulate MEK activity and (ii) one or multiple additional downstream effectors control the EPI/HYPO genetic program. For example, platelet-derived growth factor (PDGF) signaling is critical in regulating HYPO cell survival through the PI3K-mTOR pathway in mice (Artus et al. 2010, Bessonnard et al. 2019).
Functional interactions at work between the three main blastocyst tissues during elongation
The extraembryonic and embryonic tissues that make up the blastocyst are frequently defined according to their future fate, by their eventual contributions to the placenta and extraembryonic membranes or to the embryo proper. However, much experimental evidence, old and new, suggests that these tissues also have functional interactions during preimplantation development, in particular to harmonize and synchronize their reciprocal development and growth.
It is likely that these interactions are necessary for the proper development of each of the three main tissues that make up the blastocyst.
Experimental evidence from trophoblastic vesicles To our knowledge, only a few studies have demonstrated the importance of interactions between extraembryonic and embryonic tissues for conceptus elongation and survival in ungulates. The earliest studies were carried out using novel (at the time) cellular tools, the trophoblastic vesicles (TV) developed by Gardner and Johnson (1972).
TVs consist of fragments of blastocysts that are devoid of EPI cells. Depending on the embryonic stage from which they are derived, TVs can be composed either of TB cells only or both TB and parietal hypoblast (PH) cells. TVs can be cultured in vitro and transferred into synchronized recipient females. In a pioneering study, Surani and Barton (1977) delayed implantation by
transferring murine blastocysts or TVs into progesterone- treated ovariectomized females. In that environment, the embryos remained in a period of quiescence that resembled diapause. Following estradiol injection, both blastocysts and TVs were able to resume implantation, indicating that implantation does not require the presence of an ICM. Interestingly, while the number of cells increased in quiescent embryos from 60 to 120, this number remained unchanged in TVs, an indication that paracrine signals from the ICM regulate TB cell proliferation. Likewise, co-culture of single bovine blastocysts with TVs improved their development in vitro, further evidence of the existence of paracrine signals between the TB and ICM (Camous et al. 1984, Pool et al. 1988, Mori et al. 2012).
In the ovine model, TVs can be maintained in vitro for up to 20 days, but do not proliferate or elongate (Fléchon et al. 1986). However, they do elongate when transferred into recipient females, but to a lesser extent than control embryos. These data demonstrate that the embryonic disc is not necessary for TB elongation but suggest again that paracrine signals and direct interactions between the embryonic disc and extra-embryonic parietal tissues may be important for the survival, coordinated growth, and morphogenesis of the conceptus during elongation (see also Hue et al. 2007).
In elongating pig embryos, several paracrine signals have been identified between the EPI and TB, and involve both the FGF and BMP signaling pathways (Valdez Magaña et al. 2014). EPI cells express FGF4, which acts through FGFR2 on adjacent mural TB cells to activate MAPK phosphorylation. In parallel, BMP4 is produced by the extraembryonic mesoderm and HYPO cells and interacts with BMPR2 expressed in TB cells. These molecular dialogs between embryonic and extraembryonic tissues appear to be evolutionarily conserved, as they play important roles also in early embryogenesis in the mouse (Goldin & Papaioannou 2003, Graham et al. 2014, Kurowski et al. 2019).
However, several aspects still remain to be clarified in ungulates, including (i) the precise origin of the paracrine signals required during elongation, (ii) the mechanisms of molecule delivery, and (iii) the effects of these signals on their target cells.
The use of trophoblastic vesicles may help to answer these questions if future studies can take into consideration their cell composition. Indeed, in previous studies, TVs derived either from equine (Ball et al. 1989) or ovine blastocysts (Fléchon et al. 1986) were composed of numerous cell types that have not been fully characterized. But another complementary strategy could be the use of medium- and high- throughput transcriptomic technologies, which have already shed some light on various aspects of ungulate preimplantation development.
Evidence from early transcriptomic studies:
differences between developmental stages
An indispensable tool in elucidating differences among developmental stages has been the use of transcriptomic profiling, which has helped to identify important cellular processes triggered in TB and HYPO during elongation.
In sheep and cattle, comparisons of gene expression between ovoid, tubular, and early filamentous stages confirmed that a key molecular transition occurs between the ovoid and tubular stages with the initiation of elongation (Cammas et al. 2005, Degrelle et al. 2005).
By integrating multiple independent transcriptomic studies of different stages of elongation (Ushizawa et al.
2004, Hue et al. 2007, Mamo et al. 2011), researchers were able to identify key differentially expressed genes (DEGs) whose functions are linked to cell proliferation, cellular growth and differentiation and connective tissue formation (Hue et al. 2012). Interestingly, one recurring function of trophoblast DEGs is an association with lipid metabolism. This particularity of ungulate blastocysts, which has not been reported from mice or humans, seems to be linked to the fact that the elongation process in ungulates is coupled with a long preimplantation period. During this time, the conceptus relies only on its own resources and uterine histotroph to meet its significant resource demands for cell growth and cell proliferation. For example, for bovine blastocysts, the wet weight of the conceptus can increase by a factor of 20 between gestational days 16 and 19 (Lewis et al. 1982). A recent transcriptomic comparison between ovoid, tubular, and filamentous bovine blastocysts clearly highlighted important changes in lipid metabolism and fatty acid biosynthesis during elongation, together with alterations in arachidonic acid metabolism and prostaglandin synthesis and transport (Ribeiro et al. 2016a). This confirms the importance of lipids as a key element during early embryogenesis for many aspects of cell biology, including cell growth and phospholipid membrane synthesis, energy production, and intercellular signaling (see review by Ribeiro et al.
2016b). Similar observations have been made in ovine or porcine conceptuses undergoing elongation (Charpigny et al. 1997, Blomberg et al. 2006, Waclawik et al.
2013, Brooks et al. 2015). In ruminants, the regulatory factor PPARG (peroxisome proliferator-activated receptor gamma) has proven to be an integral aspect of this process: it plays a significant role in governing prostaglandin synthesis and lipid metabolism and is required for conceptus elongation (Brooks et al. 2015, Ribeiro et al. 2016b) and likely for TB differentiation as well (Degrelle et al. 2011). A recent gene expression analysis presented evidence for a similar function for PPARG in the extra-embryonic layers in pigs (Blitek &
Szymanska 2019).
An important limitation of these studies is that they compare gene expression profiles acquired from whole
embryos. This provides functional information on the molecular dynamics among developmental stages but does not clarify precisely the specific functions and connections between the cell populations that make up the conceptus. One way to better answer this question would be to isolate each subpopulation and analyze its transcriptome profile independently or, better yet, to characterize the blastocyst at the single-cell level during the preimplantation period.
Evidence from dissection-based transcriptomic studies:
compartmentalization of layer functions
Using dissection-based strategies multiple studies have recently performed transcriptomic analyses of equine (Iqbal et al. 2014), pig (Bernardo et al. 2018), and bovine embryos (Hosseini et al. 2015, Zhao et al. 2016, Pfeffer et al. 2017, Bernardo et al. 2018). These have provided new insights into the signaling pathways that are potentially active in the blastocysts of these three species. Specifically, Iqbal et al. (2014) compared the transcriptome of ICM and TB using bisected D7 equine blastocysts, while Bernardo et al. (2018) characterized the transcriptomes of the EPI from three mammalian species (mouse, cattle, pigs) at three equivalent stages:
ICM, early epiblast, and late epiblast. Hosseini et al.
(2015) characterized ICM and TB isolated from in vivo- produced D7.5 expanded bovine blastocysts, while Zhao et al. (2016) used immunopurified TB and ICM cells from in vitro-produced D7 bovine blastocyst and Pfeffer et al.
(2017) compared the transcriptomes of the late epiblast and the embryonic disc (including early epiblast and the underlying hypoblast) of bovine embryos.
The latter studies used the mouse model as a reference for identifying developmental stages and tissues. Taken together, these studies enable us for the first time to (1) identify shared and species-specific molecular players that govern pluripotency at these three stages and (2) highlight the differences between mice and ungulates with respect to pluripotent tissues. By looking at orthologous genes shared among the different species, Bernardo et al. (2018) first observed that the mouse EPI (regardless of the stage under examination) is quite divergent from that of pigs or cattle, which are more similar to each other transcriptionally. A similar observation was made by Hosseini et al. (2015) in a comparison of the transcriptomic profile of bovine ICM to that of humans and mice. In addition, Bernardo et al. observed that the three species’ ICM profiles were distinct from those of early and late epiblasts, but differences between ICM and EPI were much more pronounced in mice than in ungulates. These results suggest two intriguing hypotheses: that the current definition of naive and primed states of pluripotency could merely represent a particularity of rodents or that naive pluripotency could be a very labile and transient state in ungulates. Both
studies also confirmed the molecular similarity between early and late epiblast, supporting the existence of a stable and unique primed-like pluripotent state that is maintained during elongation in equine, pig, and cattle embryos. Indeed, most of the molecular players known to regulate primed pluripotency in the EPI of mammalian embryos are expressed in the embryonic disc of pig and bovine embryos.
NODAL signaling, which is thought to sustain primed pluripotency in humans and mice (Vallier et al. 2005, Brons et al. 2007), is active in both bovine and pig blastocysts (Blomberg et al. 2008, Alberio et al. 2010, Hosseini et al. 2015). The studies highlighted above confirmed the expression of GDF3 in the ICM of horse embryos (Iqbal et al. 2014) and the strong expression of NODAL and GDF3, together with their receptors and transducers, in the bovine embryonic disc (EPI + HYPO) (Pfeffer et al. 2017) and in pig EPI cells (Bernardo et al.
2018). Instead, genes necessary for the JAK/STAT3 pathway, which has been described as a key feature of naive pluripotency (Hall et al. 2009, van Oosten et al.
2012), are downregulated between the ICM and early epiblast stages (Eckert & Niemann 1998, Alberio et al.
2010, de Ruijter-Villani et al. 2015, Pfeffer et al. 2017, Bernardo et al. 2018). The same was also observed for WNT signaling, which is known to maintain naive pluripotency by inducing the nuclear translocation of TCF3 and is apparently inactive in the EPI of pig and bovine blastocysts (Pfeffer et al. 2017, Bogliotti et al.
2018) but active in bovine TB (Zhao et al. 2016).
Interestingly, BMP signaling is also active in EPI of pig, equine and bovine embryos. In mice, BMP signaling has been shown to support the maintenance of pluripotency through the activation of Id1 and Id3 (Ying et al. 2003). In ungulates, BMP4 is expressed in EPI at all stages, but with notable variability. BMP2 is also strongly expressed in the ICM of equine blastocysts (Iqbal et al. 2014) and in the embryonic disc at the ovoid stage in pigs (Valdez Magaña et al. 2014) and cattle (Pfeffer et al. 2017). These results suggest a synergistic action of these two molecules to regulate the balance between pluripotency and differentiation of embryonic cells in mice and ungulates. Altogether, although many questions remain, the work that has been done to date confirms the biological importance of NODAL and BMP signaling pathways in the primed epiblast in mammals.
Looking beyond the embryonic disc, Pfeffer et al.
(2017) tried to analyze the flow of molecular information between embryonic and extra-embryonic tissues by dissecting the four main tissues of the bovine ovoid blastocyst (embryonic stage 5 at D14 post fertilization, Van Leeuwen et al. 2015) and comparing their transcriptomic profiles. Pathway analysis confirmed the functional proximity of HYPO and EPI within the embryonic disc, while the parietal hypoblast (PH) was less similar and occupied an intermediate position.
The transcriptome of the TB, instead, was quite unique,
demonstrating specific enrichment for biological functions linked with steroid biosynthesis. This was consistent with an earlier study of equine blastocysts, in which genes that were overexpressed in TB with respect to other tissues were specifically associated with biological processes related to lipid biosynthesis, ion transport, and Golgi vesicle transport (Iqbal et al. 2014).
By performing a detailed analysis of the genes encoding signaling molecules and their respective receptors, Pfeffer et al. (2017) confirmed that the four tissues should be theoretically able to transactivate each of the signaling pathways analyzed. Indeed, while the bovine TB expressed only a few signaling molecules (including FGF2, PDGFA, and vascular-endothelial growth factor VEGFA/VEGFB), molecules secreted by either the underlying HYPO (FGF10, Indian Hedgehog, Insulin- like growth factor IGF2, Angiopoietin, or WNT11) or the adjacent embryonic disc (NODAL, GDF3, or BMPs) may also act on TB through a paracrine action (Fig. 2).
Unfortunately, the challenges inherent in performing microdissection – namely, the limited amount of biological material obtained and possible contamination with adjacent tissues – have limited our ability to produce an accurate, dynamic molecular atlas of the developing blastocyst.
Evidence from single-cell studies
Recent developments in single-cell transcriptomics have opened new dimensions through the resolution of spatial cellular heterogeneity and the ability to capture cellular changes over time. Single-cell transcriptomes from whole embryos and tissues could help to identify cell–cell interactions, such as those between fetal and maternal cells during placentation in humans (Vento- Tormo et al. 2018) or to clarify cell lineages and inter- cellular signaling (Deng et al. 2014, Petropoulos et al.
2016, Mohammed et al. 2017, Rivron et al. 2018, Stirparo et al. 2018). Indeed, by combining blastoid culture and single-cell transcriptomic studies, Rivron et al. (2018) were able to describe the molecular exchange between TB, HYPO and EPI of mouse blastocysts. They showed that paracrine signals emitted by the EPI, including IL11, FGF4, BMP4, and NODAL together with autocrine signals such as IL11, WNT6, and WNT7B, are necessary for TB development and in utero implantation.
To date, only a few single-cell gene expression studies have been performed on ungulate species, and only on pig and bovine blastocysts (Negrón-Pérez et al. 2017, Wei et al. 2017, 2018, Ramos-Ibeas et al. 2019).
In ruminants, two independent groups recently reported single-cell qPCR analyses on in vitro-produced blastocysts (Negrón-Pérez et al. 2017, Wei et al. 2017).
Both studies used a set of genes known to be linked in mammals with early embryonic development (cell fate, pluripotency, signaling pathways). They isolated, respectively, 96 and 67 cells from the morula stage to
the expanded blastocyst. From these datasets, they were able to discriminate among the three main cellular populations of the blastocyst. While the bovine EPI expressed known markers of core pluripotency from mice and humans (NANOG, SOX2, POU5F1, NR5A2), most of the prototypical hypoblast markers were not hypoblast-specific in bovine blastocysts and were also expressed either in TB cells (GATA6), in EPI (GSC and HNF4A), or in both (FN1). Markers that were indeed largely hypoblast specific included SOX17 and GATA4, together with PDGFRA. Bovine TB cells specifically expressed or overexpressed TB-specific transcription factors known from humans and mice (CDX2, GATA3, GRHL1, GRHL2, MSX2, TFAP2A) and genes that control cell–cell interactions (KRT8, PECAM1, DAB2, ATP12A).
While these two studies helped to identify blastomere- specific gene expression and early lineage specification in bovine blastocysts, the numbers of genes and cells analyzed (a total of 17 cells for epiblast, 51 for hypoblast, and 95 for TE) were too limited to go deeper into the molecular interactions among these three tissues.
In pigs, two studies also addressed similar questions, one with single-cell qPCR on in vitro-produced pig blastocysts (Wei et al. 2018) and the other with mRNA- seq sequencing on a panel of cells isolated from in vivo- derived pig conceptuses at several developmental stages (morula, early blastocyst, late blastocyst (hatched), and spherical blastocyst; Ramos-Ibeas et al. 2019). The first interesting comparison to be made from these two publications is the difference between blastocysts produced in vitro and those produced in vivo. Among the blastocysts produced in vitro, there was no clear
transcriptomic segregation between HYPO and EPI even at the late blastocyst stages; instead, segregation was quite evident between hypoblast and epiblast in late blastocysts produced in vivo. These results provide evidence that in vitro-produced blastocysts 8 days after fertilization are not as mature as those developed in vivo and that culture conditions seem to be inadequate at this stage.
The two studies confirmed the presence of specific markers for TB (GATA2, GATA3, DAB2, CDX2) and EPI (POU5F1, SOX2, NANOG, KLF4) that are conserved among mammals in early blastocysts. In addition, Ramos- Ibeas et al. reported specific markers for the pig hypoblast in hatched and spherical blastocysts (PDGFRA, GATA4, GATA6, COL4A1, NID2, RSPO3). The genes POU5F1, KLF4, GATA6, and PDGFRA were also expressed in the morula and ICM of early blastocysts, which supports the shared origin of HYPO and EPI cells from the ICM.
This latter study also investigated the signaling pathways active in the different cell layers and their importance for cell commitment to different lineages. To do this, they first performed differential gene expression analysis between cell groups to identify DEGs associated with known signaling pathways. Then, they cultured pig embryos with inhibitors of these previously identified candidate pathways. Using this two-step approach, they observed a switch in the dependence on JAK/STAT, PI3K, and TGFβ pathways for the maintenance of pluripotency in epiblast cells. While JAK/STAT and PI3K seem to be necessary at the morula and early blastocyst stages, NODAL signaling is preponderant later in the EPI. PI3K also seems to be important for TB development in early
Figure 2 Representation of molecular flows among different embryonic, extra-embryonic, and maternal tissues. AAs, amino acids; EPI, epiblast;
IFNs, interferons; PGs, prostaglandins; PH, parietal hypoblast; RBPs, retinol-binding proteins; TB, trophoblast; SLC, solute carrier transporters;
UT, uterus; VH, visceral hypoblast.
blastocysts. Unlike reports from cattle, these authors did not observe a particular effect of WNT inhibition on these cells but they did confirm the dependence on MAPK signaling in the ICM for the formation of HYPO (see also ‘EPI/HYPO specification’ section).
This study did not pay particular attention to peroxisome proliferator-activated receptor (PPAR) signaling, even if the DEG analysis clearly highlighted its activity in both the hypoblast and epiblast of hatched and spherical blastocysts (Ramos-Ibeas et al. 2019).
PPARs are nuclear receptor proteins that are able to dimerize with RXR to control gene expression; they are particularly relevant because their main ligands are free fatty acids and eicosanoids, of which the latter group is produced by the oxidation of arachidonic acid or other polyunsaturated fatty acids and includes prostaglandins and leukotriens. The importance of prostaglandins during elongation has been extensively documented in ungulates (Dorniak et al. 2011, Brooks et al. 2015). Together with gene expression studies in cattle (Ribeiro et al. 2016, Pfeffer et al. 2017), this work confirms the importance of the embryonic and parietal extra-embryonic tissues (HYPO and TB) for the synthesis of PPAR ligands from fatty acids present in the uterine fluids.
Taken together, all the studies discussed here highlight the interdependence that exists among the tissues of the conceptus (TB, EPI, HYPO) to ensure its development and survival. Importantly, these studies also suggest that these tissues, through the production and secretion of many different molecules (Table 1), interact strongly with the maternal uterine environment (ovary and endometrium), constituting an inseparable ‘ménage à quatre’.
Maternal influence on conceptus development Evidence of maternal effects on
conceptus development
Current limits of embryo culture
Efforts to optimize in vitro embryo production and reduce differences between in vitro- and in vivo- produced embryos in many different species (mouse, human, rabbit, cattle, pig, sheep) have led to a deep appreciation for the extent of the maternal influence on conceptus development. While the quality of oocytes and the conditions of in vitro maturation or fertilization undoubtedly matter (Smith et al. 2009, Leroy et al. 2015), the maternal influence on the developmental phase that precedes implantation and placental formation is unmistakable and appears to derive mostly from the molecular crosstalk at work within the maternal tract.
This effect arises first in the oviduct (from one cell to early blastocyst stage, especially in horses due to a longer stay therein, Smits et al. 2016), and then moves into the uterus (from early blastocyst to implantation).
For decades, efforts have been made to establish in vitro systems that mimic the oviduct environment to produce blastocysts from different species (Smits et al.
2012, Fowler et al. 2018, Hamdi et al. 2018). Instead, in vitro development of older stages has only succeeded in human and mice through cultures in enriched media (Hsu 1973) or, recently, the establishment of embryoids/
gastruloids (Govindasamy et al. 2019). In livestock species, embryonic discs have been successfully cultured for a few days with no effect on gastrulation patterns but extra-embryonic tissues have failed to elongate in vitro despite their good in vitro survival in most species (Hochereau-de Reviers & Perreau 1993, Wianny et al.
1997, Valdez Magaña et al. 2014, Stankova et al. 2015, see also ‘Experimental evidence from trophoblastic vesicles’ section).
Diapause: when the maternal environment controls the onset of blastocyst development
Scientific interest began to focus on the in vivo maternal influence on embryonic tissues following reports that mouse embryonic stem (ES) cells were more easily derived from females in diapause (Evans & Kaufman 1981, Kaufman et al. 1983), a state of embryonic dormancy controlled by signals from the uterus (Renfree
& Fenelon 2017). In ungulates, the roe deer is the only species in which naturally occurring embryonic diapause has been studied. During this diapause, which lasts for 5 months, the TB and the EPI exhibit a minimum proliferation rate, do not differentiate, and display unique structural features including a lack of mitochondria, ribosomes, Golgi apparatus, and endoplasmic reticulum (Aitken et al. 1975). In parallel, the endometrium also presents specific and dynamic ultrastructures, indicating that endometrial glands also play a major role in diapause induction and the subsequent reactivation of embryonic development (Aitken et al. 1975). Recently, characterizations of the proteome of uterine fluids at different times during and after diapause highlighted the importance of polyamine biogenesis and degradation during roe deer diapause and the importance of the regulation of cell–cell adhesion during this process (van der Weijden et al. 2019). Hormonal regulation is also associated with exit from diapause in roe deer, but this remains poorly characterized; data from the existing literature are contradictory regarding levels of estrogen, progesterone, and prolactin during and after diapause (Aitken 1974, Lambert et al. 2001, Korzekwa et al. 2019, van der Weijden et al. 2019).
Natural diapause serves as a clear indicator of the importance of the maternal environment in controlling the timing of embryonic development before implantation. This observation has also been confirmed with experimental data. Diapause has been induced by (i) the transfer of embryos from a non- diapausing species to the uterus of a diapausing species
Table 1Biological molecules known to be produced by the different embryonic, extra-embryonic and maternal tissues in different ungulate species. Growth factors and cytokinesHormonesAmino acidsFatty acidsProstaglandins
Extracellular
vesicules nucleic acids
ReceptorsCarbohydrates Embryonic Disc EpiblastBMP4[p,c], PDGFA[p,c], FGF2[c,h], WNTinh[p,c], WNT11[p,c], NODAL[p,c], GDF3[p,c,h], FGF19[p], FGF4[p,c,h]
Polyamines[p,s]PPARs[p,s,c]Proteins, Lipids, miRNAs[p,s,c,h]FGFRs[p,c,h], PDGFRA[p,c,h], BMPRs[p,c,h], TGFR[p,c,s], HGFR[c], IGFR[c], VEGFR[c], WNTRs[c] HypoblastBMP4[p,c], PDGFA[p,c], FGF2[c,h], WNTinh[p,c], WNT11[p,c], BMP2[p,c,h], BMP7[p,c]
E2[p]Polyamines[p,s]PPARs[p,s,c]Proteins, Lipids, miRNAs[p,s,c,h]FGFRs[p,c], PDGFRA[p,c,h], HGFR[c], IGFR[c], VEGFR[c], WNTRs[c] Parietal extra-embryonic layers HypoblastIHH[c], IGFs[c,h], WNT11[c] FGF10[c], FGF2[c]
E2[p,c,h], Placental Lactogene[s]Polyamines[p,s]PPARs[p,s,c], SLCs[p,s,c,h], Cortisol[s], Arachidonate[p,s,c], RBPs[p,s,c,h]
LPA[p,s,c], LPARs[p,s,c], PGE2[p,s,c,h], PGF2α[p,s,c], PGI[p]
Proteins, Lipids, miRNAs[p,s,c,h]FGFR[p,c]Fructose[p,c,s] TrophoblastVEGF[c], FGF2[c], PDGFA[c], IFN-τ[c,s], IFN-ϒ[p], IFN-δ[p] IL1B2[p], TNFα[h]
E2[p,c,h], Placental Lactogene[s]Polyamines[p,s]PPARs[p,s,c], SLCs[p,s,c,h], Cortisol[s], Arachidonate[p,s,c], RBPs[p,s,c,h]
LPA[p,s,c], LPARs[p,s,c], PGE2[p,s,c,h], PGF2α[p,s,c], PGI[p]
Proteins, Lipids, miRNAs[p,s,c,h]FGFR[p,c], BMPR[p,c], PTGER[p,s], PTGIR[p,s], LPAR[p,s,c], TGFR[p,c,s]
Fructose[p,c,s] Endometrium Uterine glandsFGF7[p,s], FGF10[s], EGF[p], TGFs[p], IL6[p], FGF2[p], IGF[p,s], WNTs[s,c], GH[s], CSF2[p,c,s], HGF[s], SPP1[p,s]
AAs[p,c,s,h], SLCs[p,c,s] Polyamines[,s]
RBPs[p,s,c,h], Retinol[p,s,c], SLCs[p,s,c,h], Arachidonate[p,s,c]
LPA[p,s,c], LPARs[p,s,c], PGE2[p,s,c], PGF2α[p,s,c,h]
Proteins, Lipids, miRNAs[p,s,c,h]IL1R[p,s,c], LPAR[p,s,c], IFN-R[p,s,c], FGFR[p,s,c] ESRs[p,s,c],PGR[p,c,s,h], HGFR[s], TGFR[p,s,c], PTGIR[p,s]
Glucose, fructose, mannitol/ sorbitol
[p,c,h,s] BloodIGF1[p,c,h,s]P4[p,c,h,s]NEFAGlucose, fructose, mannitol/ sorbitol
[p,c,h,s] c, cattle; p, pig; h, horse; s, sheep.
(ferret embryo to mink uterus; sheep embryo to mouse uterus) or (ii) the addition of uterine luminal fluids (ULF) from a diapausing species to a culture of blastocysts from a non-diapausing species (mouse ULF to rabbit embryos;
wallaby ULF to non-diapausing mouse embryos) (Ptak et al. 2012, Fenelon et al. 2017). Reactivation of diapaused embryos is inducible as well through the use of ULF, co-culture with uterine cells (mink), or transfer to a reactivated uterus. So far, the uterine signals that reactivate development are understood better than those that induce diapause (Renfree & Fenelon 2017), but the control of polyamine synthesis seems to be a key conserved factor for the induction and maintenance of and exit from diapause (Lefevre et al. 2011, Fenelon et al. 2017, van der Weijden et al. 2019). However, the maternal factors that control diapause and embryonic receptivity to these signals might not be universally identical: pig embryos transferred to diapaused rat uteri degenerated, whereas sheep embryos transferred to diapaused mouse uteri did not, and furthermore, some were even able to elongate when transferred back to sheep uteri (Ptak et al. 2012, Geisert et al. 2017).
Experimental evidence in vivo and ex vivo
So far, the maternal influence on extra-embryonic tissues in vivo has been described through its effects on the regulation of elongation processes in ungulates or on capsule formation in equids. Blastocysts as well as trophoblastic vesicles do not elongate in vitro but do so in vivo once transferred to the uteri of synchronized recipients (Heyman et al. 1984, Fléchon et al. 1986).
When there is a reduced density of glands in the uterus, or no glands at all (such as in the sheep uterine gland knockout model UGKO), elongation is reduced or abolished (Gray et al. 2001, Spencer & Gray 2006).
Scientific efforts have thus focused primarily on uterine secretions, but a few studies have examined physical constraints on embryo elongation. Cultures of bovine and porcine embryos using capillaries or hydrogels demonstrate only limited elongation in vitro (Vejlsted et al. 2006, Hue et al. 2007, Miles et al.
2017) and equine embryos, which do not elongate but instead expand into spherical vesicles, are protected by a capsule that needs the uterus to form and can withstand myometrial forces (Quinn et al. 2007, Stout 2016). Beside the role of physical forces, another open question is whether uterine secretions act directly or indirectly on extra-embryonic tissues. In pigs, priming from mesoderm cells is needed through BMP signaling before trophoblast cells can respond to specific uterine signals like FGF signaling (Valdez Magaña et al. 2014).
Nonetheless, recent work has indicated that uterine secretions have many more functions beyond those involved in developmental progress, including the coordination of on-time implantation (mouse) or stromal decidualization (Kelleher et al. 2018, 2019).
Interplay between maternal and conceptus tissues to drive late blastocyst development
During pregnancy, maternal uterine secretions (or histotroph) contain a variety of proteins, amino acids, carbohydrates, lipids, ions, and extracellular vesicles, which are produced by the uterine glandular epithelium (GE) and the luminal epithelium (LE) or transported from the serum to the uterine cavity by the luminal epithelium. Developing conceptuses, instead, produce signals for MRP (maternal recognition of pregnancy), stimulate endometrial receptivity, and secrete proteins, lipids, and extracellular vesicles that contribute to the ULF. Together, the multitude of factors involved in communication between maternal tissues and those of the conceptus contribute to create an interplay of formidable temporal and spatial complexity. Below, we focus on the secretions from both sets of tissues and their consequences for embryonic and extra- embryonic development.
Uterine secretions
Studies of the metabolome and proteome of uterine fluids have been reported for several species, including cattle, horses, pigs, and sheep (Li et al. 2007, Forde et al.
2014a,b, Bastos et al. 2019, reviewed in Spencer et al.
2019). In addition, transcriptomic studies have revealed specific secretory functions for the epithelial and stromal components of the uterus (Bauersachs & Wolf 2012, Zeng et al. 2018). These secretions result from the action of regulatory pathways on the endometrium that initiate with priming of ovarian origin (estrogens then progesterone). Additional stimulations, such as those from the conceptus, can also modulate the composition of these secretions (Groebner et al. 2011, Gibson et al.
2017), together with other environmental factors like the physiological status (Satterfield et al. 2010, Forde et al. 2014a, Beyer et al. 2019) and the diet of the mother (Chartrand et al. 2003, Giller et al. 2018, Crouse et al. 2019).
The release of histotroph components in the ULF can occur either by direct secretion or in extracellular vesicles. Extracellular vesicles (originating from both LE and GE) contain a multitude of mRNAs, miRNAs, proteins, and lipids, as well as surface receptors/ligands, with a composition that varies based on cell type (Burns et al. 2014, 2016). The number and content of maternal extracellular vesicles are controlled with progesterone (Burns et al. 2018) but extracellular vesicles can also be produced by the conceptus itself and act either on conceptus cells, to facilitate implantation (Desrochers et al. 2016), or on endometrial cells (LE, GE; Kusuma et al. 2016, Nakamura et al. 2016).
Amino acids and carbohydrates are essential components for the development of early embryos (Booth et al. 2005, Thompson et al. 2016) and fetuses (Wu et al.
2008), as illustrated by the infertile phenotype of UGKO
sheep (Spencer et al. 2019). During early pregnancy in ewes, there is a marked increase in the abundance of most amino acids, including arginine (Arg), leucine (Leu), glutamine (Gln), glutamic acid (Glu), methionine (Met), serine (Ser), and histidine (His), as well as glucose, calcium, and sodium (Bazer et al. 2015), reflecting their importance for survival and growth of the conceptus.
Interestingly, in addition to glucose, fructose is the most abundant hexose sugar in the uterine fluids of ungulate mammals, but its role in the growth and development of the conceptus remains imperfectly understood (Kim et al. 2012). The composition of amino acids and carbohydrates in ULF also reflects the expression levels of their respective transporters in endometrial and conceptus cells, as shown in studies of cattle and pigs (Forde et al. 2014b, Bazer et al. 2015, Steinhauser et al. 2016).
As an example, concentrations of Arg and glucose in ULF of ewes increase significantly between D10 and D15 of pregnancy; these molecules are transported from maternal blood to uterine lumen through the activity of specialized transporters, SLC2A1 and SLC5A1 for glucose and SLC7A2B for arginine. Of these, the expression of SLC5A1 is induced in the endometrium (LE, GE) by progesterone, while SLC2A1 and SLC5A11 are induced by P4 and IFN-tau (Gao et al. 2009a). Similarly, in pigs, luminal cells express AKR1B1 and SORD, two enzymes that are necessary to convert glucose into fructose, and endometrial cells also express the fructose transporter SLC2A8 during the peri-implantation period (Steinhauser et al. 2016). Similar observations have been reported for other amino-acid transporters (Gao et al.
2009b,c, Satterfield et al. 2010, Bazer et al. 2015) and nutrient-sensing pathways (Gao et al. 2009d). As a result, the concentration of amino acids and carbohydrates varies in time in the ULF during early pregnancy and is dependent on the expression of transporters in the conceptus and/or uterus (Gao et al. 2009a, Forde et al.
2014a, Bazer et al. 2015, Gibson et al. 2018).
As a consequence of the increase in available amino acids, polyamines – a specific product of amino- acid biosynthesis – are also detected in significant quantities in ULF. Polyamine synthesis required Arg, Pro, and L-Ornithine together with the enzyme ornithine decarboxylase (ODC1), leading to the production of putrescine, spermidine or spermine (Lefevre et al. 2011).
Polyamine levels in ULF are probably dynamically regulated by a balance between biosynthesis and catabolism and by transport between intra- and extra-cellular environments (Persson 2009). Indeed, fluctuations in polyamine levels during diapause in roe deer and mice have been linked with increased expression of ODC1 and SMS (spermine synthase) in the luminal stroma (Zhao et al. 2012, van der Weijden et al. 2019).
Histotroph also contains lipids and lipid mediators such as prostaglandins (PGs) or lysophosphatidic
acid (LPA). In ruminants, retinol and lipids have been detected in uterine fluids during diestrus, whereas studies of the conceptus at the onset of elongation have noted transporters and binding proteins for fatty acids (SLCs) or retinol (RBP4) (Ulbrich et al. 2009, Liszewska et al. 2012, Mullen et al. 2012). In pigs and horses, the histotroph also contains small-lipid-transport proteins such as RBP, uterocalin, and uteroglobin, with the last delivering lipids to the conceptus via the hydrophilic capsule glycan (Stallings-Mann et al. 1993, Suire et al. 2001, Quinn et al. 2007, Waclawik et al. 2013, Jeong et al. 2016). The active transfer of lipids from the endometrium to the conceptus was proposed decades ago by Boshier et al. (1987) but has only recently been supported by studies documenting extracellular vesicle formation and uptake (Mulcahy et al. 2014, Mathieu et al. 2019).
Conceptus secretions
Through its secretions, the conceptus contributes to the MRP (which mostly relies on the inhibition of luteolysis), but also prepares the intimate cellular dialogue between the TB and the endometrium that is necessary for implantation (Vento-Tormo et al. 2018, Biase et al. 2019).
The production of prostaglandins by TB appears to be indispensable for this purpose in many ungulate species (Brooks et al. 2014), while other molecules appear to be more species specific, such as estrogens and IL1B2 in pigs (Ka et al. 2018) or IFN-tau in ruminants.
IFN-tau is produced by the ruminant conceptus and is known for its combined effects on the endometrium as well as on extra-uterine cells or tissues including PBMC, liver, and corpus luteum (Hansen et al. 2017, Imakawa et al. 2018, Passaro et al. 2018). In pregnant sheep, IFN-tau loss-of-function experiments using anti- sense oligonucleotides resulted in embryonic growth retardation and malformations (Brooks & Spencer 2015).
Strikingly, loss of the type I IFN receptor has no effect on embryonic elongation (Brooks & Spencer 2015), which suggests that IFN-tau may act indirectly on conceptus elongation, potentially through its direct effects on endometrial cells.
Although IFN-tau has not been detected in non- ruminant ungulates, other mediators for the recognition of pregnancy have been identified in pigs. From the ovoid stage, the porcine blastocyst produces estrogens and IL1B2 (Perry et al. 1973, Spencer et al. 2004, Mathew et al. 2015) which act synergistically to drive endometrial functions and conceptus development. IL1B2 secreted by the conceptus is necessary for elongation and endogenous estrogen production and could help inhibit luteolysis by promoting estrogen production by endometrial cells (Mathew et al. 2015, Whyte et al. 2018). However, the direct effect of conceptus- derived estrogens remains unclear. CYP19A1-null pig embryos, which do not produce endogenous estrogen,
